[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Updateref Ignored, Part Duex: The Final Chapter

>>>I'm still having some replication problems.  With kind help from
>>>Kurt Zeilenga I came to understand that it is not a good idea to
>>>have my rootdn and updatedn be the same.  So I've added a user for
>>>purposes of replication.
>>>This works well on my master slapd server.  I can now
>>>add/modify/delete users there and have changes successfully
>>>replicate the slave.  But I am still have problems at the slave
>>>end.  Basicly, when I try to add users using my new "Replicator"
>>>account on the slave the go strait to the database and ignore my
>>>updateref directive.
>>Of course.  If you are 'signed on' as the replicator, slapd assumes you
>>are the master and updates it's database accordingly.  A user 
>>should never
>>user (or really even know about) the replicator account.  Sign onto the
>>slave as a ***user*** and make changes,  and you should see the 
>Makes sense.  Tried this and it works like a champ -- thanks for 
>the help!
>I searched the list archives today (search is alive again this 
>week -- yeah!) and notice a LOT of folks approaching replication 
>problems for the first time had the same problems I've had.  So I 
>would be very happy to write everything I've learned regarding 
>replication (including working examples) and submit it all for 
>review to the OpenLDAP documentation gods.  From comments to the 
>list it seems that a lot of people read of the replication 
>documentation and still don't "get it" -- and to me the docs do 
>seem a bit sparse in this particular area.
>And sometimes a newcomer's perspective is a handy thing for writing 
>clear documentation...  :)

Have you looked at my LDAP presentation at -

I cover replication in some depth.  If you think I missed something let me 
know and I will gladly amend.

I will add a slide about the "specialness" of updatedn.

Ximian GNOME, Evolution, LTSP, and RedHat Linux + LVM & XFS